Methodology for Evaluating Model Routing Released

@tomas_hk announced the release of a methodology for evaluating model routing. The approach relies on interactive benchmarks designed to represent how agent costs accumulate. The post was retweeted by @_akhaliq, an ML researcher at Autodesk who curates AI papers on X. The announcement highlights that these benchmarks better capture cost accumulation compared to previous methods. No further details on the specific benchmarks or evaluation results appear in the provided posts.